Pure Storage Announces Cloud Data Services

Pure Storage is a current PivotNine consulting client.

In a lot of our recent work with clients, more and more people are noticing that enterprises can’t change their entire IT investment into cloud-native services overnight. You’d think this was obvious, but apparently not.

While there’s plenty of goodness in new approaches born in the cloud-native world—microservices, service composition, immutable deployments, consumption pricing, etc.—it’s not physically possible to move existing systems to these new approaches all at once. It usually involves writing new code using new architectures, and that takes time.

Successful companies already have systems that work and make them money. It’s not reasonable to ask them to turn them off while they rebuild their entire organisation around a new and rapidly changing approach. Enterprises need bridges to help them get from their existing way of doing things to the new ways, and that’s what a lot of Pure Storage’s announcements today are all about.

Pure is selling a vision of hybrid applications where data needs to be available wherever you decide to run them. This could be in your existing on-site datacentre, it could be in co-location, it could be in the cloud. It could even be out at the edge in an embedded system.

And these applications could use cloud-native approaches like microservices or functions/serverless, or it could be a well-tested and stable monolith running in a VM or on bare-metal. The actual ‘application’ might even be composed of all of these things simultaneously. If you’re interacting with a website, what do you care if the backend has components on a mainframe, a few VMs, and serverless functions in multiple clouds?

To help enterprises get to this vision of the world, Pure is proposing a data-centric architecture with the same data services available wherever you need them, controlled via the same API. To make this happen, Pure has put its software into the cloud, and called it Pure Storage Cloud Data Services.

Cloud Block Store provides high-performance storage (as you’ve come to expect from Pure Storage) in AWS with the same API you’re used to with on-site Pure appliances. This also means you can move data to-and-from the cloud to suit your needs.

Pure1 Cloud Data Management gives you a single, consistent way to manage your storage whether it’s on-site, in colo, or in the cloud. The consistent API means if you’ve invested in a particular management abstraction (perhaps vCenter, or Cloud Formation templates) for your infrastructure, it can continue to work as you extend into new ways of working. It also means that benefits gained from the new ways of working also flow back into the older systems. Investments in progress aren’t isolated to a single mode of working in an us-versus-them battle of bi-modal operations.

What Pure is describing here fits nicely into Simon Wardley’s PST model of technology evolution. Novel ideas transition into more standardised products before becoming utilities that other, new services are built upon. What’s not obvious on a 2-dimensional matrix is how novel new ideas are built on top of the utilities: grid electricity runs the datacentres the cloud lives in. EC2 is used to run Aurora. You can’t have the higher-level abstractions without the utilities underneath.

And, importantly, multiple technologies, each at a different point of its evolution, exist simultaneously. Enterprises should feel free to choose the most appropriate technology for the need they are trying to satisfy. They should build flexibility into their systems that allows things to evolve over time. Old ways are replaced with new ones constantly, when the time is right. All the components have changed, but it’s still my grandfather’s axe.

Pure is providing a way (not the only way, mind, just one way) to have storage in multiple places that looks and feels familiar no matter where you are. It gives you a common platform that you can build upon wherever you need to, and makes it easier to make changes at a higher level because the underpinnings are consistent and reliable.

If you choose and like Pure Storage, you can use it anywhere and it will feel the same. Your choice of storage subsystem won’t dictate your application design choices.

Which is as it should be.

Enterprises can concentrate on choosing the architecture that works best for them today, and know they have the flexibility to change their mind when they need to and update things to the right architecture for tomorrow.